Unit 1 · Foundational
What machine learning is, when it is the wrong tool, and the order the work happens in. Framing decisions made here determine whether anything downstream can be evaluated.
What a strong answer looks like
A strong Unit 1 answer names the task type, the target, and what success would look like numerically.
Know
Traditional programming encodes rules a person wrote; machine learning infers approximate rules from examples. Approximation is why evaluation is mandatory.
Apply
Use ML when the rules are unknown or too numerous to write, not when they are known and stable.
Watch out
Reaching for a model where a simple documented rule would be more accurate and auditable.
Study move
Name one problem better solved by a written rule and say why.
Know
A two-category outcome is binary classification, a numeric outcome is regression, and unlabelled grouping is clustering. Framing determines every later choice.
Apply
Identify the target first, then the task type it implies.
Watch out
Framing a naturally continuous outcome as classification by bucketing it, which discards information.
Study move
Classify four described problems by task type and justify each.
Know
Frame, prepare, train, evaluate, deploy, monitor. Most of the effort sits in preparation, and monitoring exists because data drifts after deployment.
Apply
Say which stage you are in when something goes wrong; the stage usually implies the fix.
Watch out
Treating deployment as the end. A model degrades as the world it was trained on changes.
Study move
Place four described activities into the correct workflow stage.
Know
A baseline is the score a trivial rule achieves. Without one there is no way to tell whether a model is contributing anything.
Apply
Compute the majority-class baseline before training anything.
Watch out
Reporting 95 percent accuracy on data that is 95 percent one class, where the baseline is identical.
Study move
For an imbalanced dataset, state the do-nothing accuracy and what would count as real improvement.
Connections and techniques that receive extra attention in this unit.
Related topics some schools attach to this unit and others leave out. Covered on request rather than assumed.
Modeling drill
For predicting whether a student submits an assignment late, state the task type, the target, the baseline, and what score would make a model worth deploying.
